@alanz @zliu41 apply-refact updates (ghc-exactprint > 1.8.0.0) will be needed

I will bump the version number on my https://github.com/alanz/ghc-exactprint/tree/ghc-9.10 branch (tomorrow) if that will help.
If I have a way of installing the new cabal for github CI, I can make a release

I will bump the version number

Done

I will bump the version number

Done

thanks alan, that is indeed a help (although of course, we will need a release in due course). @zliu41 with this cabal.project file

allow-newer: all
source-repository-package
  type: git
  location: https://github.com/alanz/ghc-exactprint.git
  tag: f334b0c2da5ca9193297e77a9d1fc620f859ffde
packages: apply-refact.cabal

cabal install all produces the errors in this pastebin https://pastebin.com/HyvFvR8R indicating a few fixes to apply-refact source itself will be needed in addition to this.

https://hackage.haskell.org/package/ghc-exactprint-1.9.0.0

cabal install all produces the errors in this pastebin https://pastebin.com/HyvFvR8R indicating a few fixes to apply-refact source itself will be needed in addition to this.

@shayne-fletcher I've fixed these errors but there are many more. I'll continue fixing them over the next week. The working branch is https://github.com/mpickering/apply-refact/tree/ghc-9.10

Note: you need to use a Cabal-3.12-based pre-release of cabal-install to be able to compile ghc-exactprint.

See https://discourse.haskell.org/t/ann-cabal-3-12-0-0-released/9504#how-to-get-the-cabal-install-pre-release-3

cabal install all produces the errors in this pastebin https://pastebin.com/HyvFvR8R indicating a few fixes to apply-refact source itself will be needed in addition to this.

@shayne-fletcher I've fixed these errors but there are many more. I'll continue fixing them over the next week. The working branch is https://github.com/mpickering/apply-refact/tree/ghc-9.10

i hope this saves you some time @zliu41:

on the ghc-9.10.1-fixes branch on my fork https://github.com/shayne-fletcher/apply-refact.git, after making this change, of the 449 tests, the failure count is reduced from 249 to 79.

@zliu41 given the magnitude of the GHC 9.10.1 upgrade, it might make sense to strip the CPP, and adjust the bounds so the next version only works with GHC 9.10.1 (as originally happened with apply-refact, IIRC).

And my notes on the changes are at https://gist.github.com/alanz/e127e7561ddf1cfeb07fbdee9a966794

given the magnitude of the GHC 9.10.1 upgrade, it might make sense to strip the CPP

@alanz Good point! Let me see how bad it is. And thanks for the notes!

Heads up all, I am in the process of making the transform API pure, should be making a new ghc-exactprint release soon.
See alanz/ghc-exactprint@5e7e8ab

https://hackage.haskell.org/package/ghc-exactprint-1.10.0.0

Update: I've made it build but there are many test failures. Will continue investigating. I don't have an ETA at the moment, but I should have some time to work on it during ZuriHac (if I haven't finished it by then).

I should have some time to work on it during ZuriHac

I can possibly lend a hand. But tied up from now until then

Update: I've made it build but there are many test failures. Will continue investigating. I don't have an ETA at the moment, but I should have some time to work on it during ZuriHac (if I haven't finished it by then).

as i mentioned above @zliu41 , removing the call to makeDeltaAst at line 739 of Internal.hs Right ast -> pure $ Right (makeDeltaAst ast)(on your branch) halves the failure count to 135 of 449 tests.

@shayne-fletcher Sorry I completely missed your message above 😅 Thanks for helping out!

Worked on it for about 2 hours today with the help of @alanz, and reduced the number of test failures from 175 to 172 😅 Will continue later.. looks like a lot more setEntryDPs needed.